Polymorphic graph attention network for Chinese NER

作者：

Highlights：

• Lexicon information has been proved to be very useful in Chinese Named Entity task.

• The existing methods make insufficient use of lexicon information.

• We propose a Polymorphic Graph Attention Network for fusion lexicon into character.

• Our method can be easily combined with pre-trained or sequence encoding model.

• The proposed method supports multi-head attention and has excellent inference speed.

摘要

•Lexicon information has been proved to be very useful in Chinese Named Entity task.•The existing methods make insufficient use of lexicon information.•We propose a Polymorphic Graph Attention Network for fusion lexicon into character.•Our method can be easily combined with pre-trained or sequence encoding model.•The proposed method supports multi-head attention and has excellent inference speed.

论文关键词：Named entity recognition,Chinese,Polymorphic Graph Attention Network,Conditional random field,Lexicon information

论文评审过程：Received 14 December 2021, Revised 27 April 2022, Accepted 28 April 2022, Available online 2 May 2022, Version of Record 13 May 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.117467